The continued rise in adolescent marijuana use and the limited effectiveness of interventions to prevent adolescent marijuana use may be due to the failure to recognize the complexity of peer relationships, including developmental shifts in the prominence of peers vs. romantic partners, changes within romantic relationships, and the interconnected relationship between peers and romantic partners, that lead to the uptake, continuation, and escalation of adolescent marijuana use. The candidate is an epidemiologist, who has been well trained in linear modeling and brings an understanding of the role of partner context on risk behaviors (i.e. condom use and concurrency). To become a successful independent investigator with a career focused on addressing the significant and complex problem of adolescent marijuana use, the candidate requires additional training on the etiology and trajectories of marijuana use and dependence, the development and impact of romantic and peer relationships on decision-making around health compromising behaviors, formal and practical training in complex System Dynamics and mathematical modeling, and a deep understanding of legal and ethical issues associated with adolescent substance use research. The long-term goals of the candidate are to become an independent investigator focused on improving adolescent health by identifying opportunities for health promotion and developmentally appropriate primary drug use prevention strategies. Marijuana use is a significant public health problem due to the breadth of health and social consequences associated with its use. Adolescent marijuana use may be an intractable problem because it is embedded in a complex system. The short-term goals of the candidate are to gain the knowledge and skills to apply a systems science approach to illuminate how the context of romantic relationships drives adolescents' uptake, continuation and escalation of marijuana use. To attain these short- and long-term goals, the training component of the award includes 4 activities: 1) develop expertise in the etiology, physiology and trajectories of adolescent and young adult marijuana abuse and dependence, 2) build a theoretical foundation of sexual, dating and romantic relationships over the course of heterosexual adolescent and young adult development, 3) gain advanced skills in systems science methods, including group model building and mathematical modeling, and 4) gain an understanding of the complex ethical issues involved in drug use research with vulnerable adolescents and young adults. The research component of the award is to quantify the influence of romantic partners on adolescent marijuana use, and to use systems science methodology to incorporate important complexities of peer socialization processes. Specific Aim 1 is to use longitudinal data to estimate the contribution of romantic partner and non-romantic peer influence on adolescent marijuana use. This knowledge will serve as the foundation for developing a quantitative System Dynamics model, which will examine whether there is an effect of romantic partners and the relative magnitude of romantic partners and non-romantic peer influence on frequency of adolescent marijuana use over time. Specific Aim 2 is to use participatory model building (e.g., convened group of multidisciplinary academic and community experts and a group of youth) to create a system diagram incorporating key components of the social system that drive an adolescent's uptake, continuation and escalation of marijuana use over time using an approach grounded in rich theory from diverse disciplines. Specific Aim 3 is to use the system diagram to develop a quantitative simulation model of peer system effects on marijuana use and calibrate the model with available data, including the empirical literature. Upon completion of this award, the candidate will become a productive investigator who has the necessary skills to conduct innovative and compelling interdisciplinary research focused on the etiology of drug use in adolescents. She will understand how social context, specifically the dynamic nature of adolescent romantic and non-romantic peer relationships, affects marijuana use. Model building may help identify feedback loops in the peer system that perpetuates disparities in health and social consequences. Findings from this work may identify leverage points in the system that may be novel, multi-disciplinary and ideally more modifiable than previously identified prevention targets.